Dive Brief:
- Nvidia is strengthening its position as a full-stack compute provider as it seeks to support agentic AI directly on devices. In partnership with Microsoft, Nvidia released RTX Spark, a chip powering Windows PCs purpose-built for personal AI agents. The company shared the news Monday during Nvidia GTC Taipei at Computex 2026.
- RTX Spark supports local AI agents and frontier models, what Nvidia Founder and CEO Jensen Huang described as a superchip. “This is the new PC,” Huang said in a press release. “The personal AI computer.”
- Nvidia also announced DGX Station for Windows, an AI supercomputer that can run frontier models of up to 1 trillion parameters locally at deskside. The product, expected in Q4, aims to move traditionally heavy enterprise AI workloads from data centers to a deskside computer that provides the necessary compute for running AI agents across Windows applications and infrastructure.
Dive Insight:
Nvidia’s move to power AI agents on devices may signal a long-term AI strategy play as the company seeks to provide compute power across the IT stack.
Enterprise adoption of AI PCs began to slow at the end of last year due to macroeconomic and trade uncertainties, according to Gartner research. In a December 2025 report, The Futurum Group also predicted a slowdown in AI PC shipment growth.
However, as enterprises begin PC refresh cycles in 2027, AI PCs are expected to replace traditional PCs, according to The Futurum Group. The global AI PC market could see a compound annual growth rate of 38% over the next five years, reaching an estimated valuation of $350 billion by 2030.
The Nvidia AI-powered PC launch in partnership with Microsoft, focused on supporting AI agents, could represent a move to position PCs as part of an enterprise’s overall AI strategy, said Ranjit Atwal, senior director analyst at Gartner.
“This is the one and only opportunity the PC market has to become part of infrastructure,” he told CIO Dive.
Nvidia likely isn’t looking to get into the PC market, Atwal said. Instead, the long-term strategy for Nvidia hinges on agent orchestration and management across devices, on premises and cloud.
“What we’re seeing here is AI agents, which are being developed, will eventually come downstream from cloud to on premises to on device,” he said. “So you’ll see them orchestrated across the board and you’ll need a platform that enables that capability across all of the hardware devices.”
